During a recent session on a friends computer where an AI agent implemented 17 Stride tasks back-to-back,I discovered something uncomfortable: the agent skipped nearly every mandatory workflow step. The task-explorer was used once, then never again. The task-reviewer was never invoked. The hook-diagnostician was never called when hooks failed. The stride-subagent-workflow skill — labeled MANDATORY — was ignored entirely. This was shocking to me as I have not see this on my own computer. I realized that I had enough contained in my configuration to force the agent to do what I wanted but not everybody had that config.
Stride Goes Multi-Platform: Copilot, Gemini, and Zero-Prompt Hook Execution
Prior to today, Stride had excellent support for Claude Code via six Skills and four SubAgents. All other agents were only provided with Skills without the enforcement and additional capabilities added by the SubAgents.
Today I shipped three updates that fundamentally change how AI agents interact with Stride. Two new platform extensions bring Stride’s full workflow and support to GitHub Copilot and Google Gemini CLI. And a new Claude Code hooks integration eliminates the most persistent friction point that still existed in the Claude Code agent workflow: permission prompts.
The future of software engineering: My Thoughts - Part 2
This is part 2 in a series where I discuss my thoughts on the Thoughtworks retreat report titled The future of software engineering. If you haven’t read part 1, I’d encourage you to do so first as it provides a lot of context.
In part 2, I will discuss the second and third categories from the report. Let’s jump right in.
2. The middle loop: a new category of work
The report states “The retreat’s strongest first-mover concept. Nobody in the industry has named this yet.” It goes on to say:
The future of software engineering: My Thoughts - Part 1
In February 2026, Thoughtworks released an article titled “The future of software engineering”. It is a great paper and I encourage everyone to read it. The article lists key themes and takeaways from attendees at a retreat they organized. Here is the Executive summary:
Executive summary
Senior engineering practitioners from major technology companies gathered for a multi-day retreat to confront the questions that matter most as AI transforms software development. The discussions covered more than twenty topics across breakout sessions, but the most significant insights didn’t emerge from one single session. Instead, they surfaced at various intersections; we found that the same concerns kept appearing in different conversations, framed by different people solving different problems.
Closing the Loop: How We Added Traceability to Stride
When AI agents review code, their findings shouldn’t vanish into the ether. That was the driving insight behind the review_report feature we shipped in Stride v1.25 and the Stride plugin v1.4 this week. Here’s the story of what we built, why it matters, and how all the pieces fit together.
The Problem
Stride’s task-reviewer SubAgent already did excellent work — it checks code changes against acceptance criteria, verifies pitfall avoidance, validates pattern compliance and alignment of the testing strategy. But the review output lived only in the agent’s conversation context. Once the agent completed the task, the structured findings were gone. Human reviewers opening a task in the Review or Done column had no visibility into what the AI had already checked.
